The purpose of this project is to study the long-term health effects of drugs, especially therapeutic agents, as they may relate to carcinogenicity. In addition, the patterns of occurrence of multiple primary cancer are evaluated in terms of implications for etiologic research. Because many studies of radiation carcinogenesis involve the evaluation of second cancers following radiotherapy for a primary cancer, it is often convenient to evaluate simultaneously the effects of chemotherapeutic agents. Populations studied include patients treated in randomized clinical trials, patients reported to cancer registries in the United States and other countries, and patients treated at several large institutions. In addition to the systematic study of therapeutic drugs, occasionally it is possible to evaluate other drug exposures in populations studied primarily for other reasons. Evaluations have been made on the effects of estrogens, tranquilizers, thyroid hormones, and anti-tuberculosis drugs on subsequent risk of cancer. A study of patients given methyl-CCNU as adjuvant therapy for gastrointestinal cancer provided the first quantitative dose-response evidence that nitrosoureas are leukemogenic in man. Alkylating agents to treat childhood cancer were associated with an increased risk of leukemia. Women with breast cancer who received chemotherapy may be at an increased risk of leukemia. Women with breast cancer who received estrogen therapy were found to have an enhanced risk of endometrial cancer. Recent concerns about the possible tumor promoting effects of thyroid supplements on breast cancer risk were not substantiated. Commonly used drugs were not found to be related to thyroid cancer, although a positive association was suggested for diuretic use.